Displacement imaging during focused ultrasound median ... · [3]. Ultrasound has been proven capable of modulating ion channel, peripheral nerve, and brain activity in a multitude

0885-3010 (c) 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TUFFC.2020.3014183, IEEETransactions on Ultrasonics, Ferroelectrics, and Frequency Control

IEEE TRANSACTIONS ON ULTRASONICS, FERROELECTRICS, AND FREQUENCY CONTROL, VOL. ??, NO. ?, AUGUST 2019 1

Displacement imaging during focused ultrasoundmedian nerve modulation: A preliminary study in

human pain sensation mitigationStephen A. Lee, Student Member, IEEE Hermes A. S. Kamimura, Member, IEEE and Elisa E. Konofagou, Senior

Member, IEEE

Abstract—Focused Ultrasound (FUS)-based viscoelastic imag-ing techniques using high-frame-rate (HFR) ultrasound to tracktissue displacement can be used for mechanistic monitoring ofFUS neuromodulation. However, a majority of techniques avoidimaging during the active push transmit (interleaved or post-pushacquisitions) to mitigate ultrasound interference, which leads tomissing temporal information of ultrasound effects when FUSis being applied. Furthermore, critical for clinical translation,use of both axial steering and real-time (




displacements within FUS pulses [27]–[30]. However, to ac-quire higher frame rates, event locking MRI acquisitions tothe FUS pulse requires multiple sonications to acquire the fulltemporal displacement profile [25]. Because neuromodulatoryeffects are unique to a specific FUS pulse, it would be moreinformative to have both spatial and temporal displacementprofiles for a single FUS pulse. Frame rates in MR-ARFImay range from 0.5 [29] to 1 Hz [28] whereas ultrasoundelastography can achieve frame rates thousands of timeshigher. Therefore, a non-invasive ultrasound-based techniquefor accurate real-time monitoring and targeting of the actualFUS beam used in neuromodulation is needed.

Ultrasound imaging techniques for tissue elasticity mea-surement have had success in various medical applications.These methods usually depend upon accurate displacementestimation of tissue between sequential ultrasound frames [31].Techniques such as acoustic radiation force imaging (ARFI),shear wave imaging, and harmonic motion imaging (HMI),utilize radiation forces to generate local displacements withinthe tissue, palpating non-invasively using ultrasound [32], [33].Tissue motion caused by this radiation force can be trackedusing cross-correlation algorithms [34]. Additionally, using 2Dor 3D probes for ARFI, the localized response in the tissuecan be tracked both spatially and temporally [35]. Feasibilityof ARFI has been demonstrated in numerous studies andin clinical applications such as breast, abdominal muscle,heart, liver, and colon imaging [36]–[38]. Acoustic radiationforce can also be used to image lesions in the tissue whereconventional ultrasound imaging has poor visualization [39],[40]. Additionally, real-time ARFI has been applied to guidesurgical procedures and high intensity focused ultrasound(HIFU) ablation [41], [42]. Since we hypothesize radiationforce is involved in the mechanism of FUS neuromodulation,we may be able to take advantage of conventional elastographyimaging techniques to monitor and target acoustic radiationforce generated by FUS pushes. We have previously demon-strated this technique in mice for mechanistic studies, relatingdisplacement to the amount of neuromodulation of the sciaticnerve [43].

As previously mentioned, conventional acoustic radiationforce imaging techniques do not image within the push pulse.For systems that use a single transducer to push and image,the contribution of all the elements are needed to remotelypalpate the tissue, thus imaging must occur after the acousticradiation force delivery. Systems using confocal arrangementsof two transducers experience interference between the pushtransducer and the imaging transducer bandwidth. Current,monitoring techniques implement filtering and/or coded excita-tion to avoid this interference [40], [44]. However, pulse lengthfor therapy and lesion monitoring indicated in these studies areon orders of seconds as opposed to




71 μs

5 ms

t

Imaging

Pulse

FUS

Push

RF Data

CUDA DAS

Notch Filtering

1D Cross Correlation

Display Displacement

226.2 222.6 227.0 227.4 227.8Time [μs]

-0.4

-0.2

0

0.2

0.4

Pre

ssu

re [

MP

a]

60 80 100 120 140 160 180-0.8

-0.4

0

0.4

0.8

Fig. 2. Within-FUS pulse sequence as recorded using a hydrophone infree-field. The FUS pulse is displayed in blue (acquired with a HFO-660hydrophone), and the imaging pulse is displayed in black (acquired with aHGL-200 hydrophone). Flow diagram of whole DAQ sequence is shown onthe right.

The power supply enables driving multiple cycle transmits.Half of the 256 connectors is connected to a 104 elementP12-5 Phased Array transducer (ATL Philips, Bothell, WA,USA) for simultaneous imaging and displacement tracking.The other 128 elements were connected to a RF matchingbox for a 4 element, 1.1 MHz FUS transducer (SonicConcepts,Bothell, WA, USA). A customized matching box directs powerfrom the 64 channels in the right connector to the 4 elementannular rings (16 mm width) with area (450, 550, 650, and70 mm2); each annular element is driven by 16 channels.The FUS transducer has an active diameter of 64 mm witha 40 mm opening where the imaging transducer was placedthrough (Fig. 1a). The attached coupling cone opening is 77mm in diameter with two tubes for degassing the water inthe transducer cone or inflation/deflation of the membrane.The focus of the FUS transducer (0.5 x 15 mm) was alignedusing a custom 3D printed mold so that the plane of theimaging transducer was centered at 30 mm in depth. TheFUS frequency and size of the focal region were chosen tooptimize adequate radiation pressure, relative to the focal size-to-nerve ratio. The focus encompasses the nerve diametercompletely in the axial direction and 20% in the lateraldirection. RF signals acquired for HFR displacement trackingwere processed in real-time using a GPU CUDA-accelerateddelay-and-sum (DAS) beamforming and 1D cross correlationalgorithm [46].

Compounded plane wave images were used for initialtargeting of the nerve. We used 5 angles (±9◦) and 1 transmit-receive operation per angle to generate a B-mode image (Fig1b) with the focus at 30 mm. Beamformed RF data was fed

into a normalized cross-correlation algorithm [46] to generatedisplacement maps of FUS pushes overlaid onto B-modeimages. A a 95% overlap and a window size of 12.25 mmwas used to track displacements in the raw RF data. Deratedpeak negative pressures used in phantoms were under 10.9MPa peak negative pressure (MI = 10.1, 72.3 W/cm2 ISPPA)and under 7.9 MPa (MI = 7.5, 29.4 W/cm2 ISPPA) in humans.

B. Simultaneous Single-system FUS Displacement TrackingPulse Sequence

Currently, the ultrasound DAQ system has not been config-ured for within-pulse imaging across two transducers withoutinterleaving sequences. Therefore, a strategy for continuousimaging and monitoring must be developed to allow displace-ment tracking within the FUS push. Simultaneous imagingof FUS pushes requires customized ultrasonic parametersfor each transducer. A 1.5 cycle plane wave emission wasprogrammed for all 128 channels connected to the P12-5. Inorder to utilize both transducers without interleaving, the pulseduration of the FUS transducer was set to a proportion of thetotal time-of-flight (TOF) for a wave to travel from the imagingtransducer face to a scatterer at the edge of the imagingwindow and back. Therefore, to generate an image at 41 mmfrom the transducer face (aperture of 9.94 mm), the TOF (71µs) sets the extended FUS burst to be an integer multiple ofthe TOF. To reach a burst pulse duration of 5 ms, matchingpulse durations seen in previous PNS neuromodulation studies,70 bursts of ultrasound without interval between pulse trainswere transmitted during the FUS push. The 71 µs bursts areprogrammed into the right 128 channels of the DAQ whichare used to drive all four annular elements.

Fig. 2 shows the pulse sequence of the proposed technique.The FUS push is shown in blue and the imaging pulse in black.The total frame rate of the technique is also depth dependent;for a depth of 41 mm, the frame-rate is 14 kHz. Hydrophonemeasurements of the FUS and imaging transducers in free field(HGL-200 & HFO-660, Onda Corp, Sunnyvale, CA) is shownin Fig. 2 middle. Hydrophone measurements reveal a 17 µsgap between subsequent FUS bursts. This was caused by acombination of electronic processing time and actuation timeof the elements. The processing sequence is defined in Fig.2. After stacking and transferring all RF imaging frames, theRF data is beamformed using a CUDA-accelerated (CUDAversion 9.1) conventional DAS beamforming. The parallelcalculations were performed using a GPU (Tesla k40c, Nvidia,Santa Clara, CA, USA) with 1024 threads and 3 dimensionalindexing. Beamformed RF data was filtered using a combnotch filter at all the fundamental (1.1 MHz) and harmonicfrequencies (2.2 - 12.1 MHz) of the FUS transducer foundwithin the P12-5 bandwidth. Harmonic Frequencies within the60% bandwidth of the P12-5 were not applied in order tomaximize signal-to-noise (SNR) of the echoes. Second orderButterworth notch bandwidths were set to 50% of the notchfrequencies and filter coefficients were calculated before theimaging sequence so that in-sequence calculations would bedevoted to beamforming and cross correlation. Displacementcalculations were performed by 1D cross-correlation with a

Authorized licensed use limited to: Columbia University Libraries. Downloaded on August 14,2020 at 09:00:42 UTC from IEEE Xplore. Restrictions apply.




95% overlap and a window size of 12.25 mm [43]. Displace-ment maps were smoothed using a 2D median filter of 0.2%of the reconstruction grid. Displacements were then overlaidonto the filtered B-mode images after calculation of all framesand displayed in real-time.

C. CUDA benchmarking

CUDA DAS beamforming was implemented through MAT-LAB GPU kernels. Variables used to beamform in real-timewere stored in the shared memory on the GPU before imagingexecution. All delays were calculated and summed up for eachpixel in the desired linear interpolation grid. 3D indexing wasused for each calculation in time (frames) and spatial (depthand lateral) points on 1024 threads divided into 3 blocks.The kernel grid size was determined as a proportion of thereconstructed frames and spatial grid. Lastly, beamformingoperation was performed by passing the raw RF data as asingular vector array.

For benchmarking GPU performance, MATLAB’s GPUparallel computing toolbox, CPU multi-core (parfor), and CPUsingle-core operations were used as comparisons. DAS usingMATLAB’s toolbox was performed by pre-allocating interpo-lated RF samples and pre-defining forwards, backwards, andcompounding delay arrays on the GPU. During imaging oper-ation, RF data was then beamformed by linearly interpolatingthe RF data onto pre-indexed and pre-defined delays. The samecalculations performed in C++ were separately performed ona single CPU and using parfor loops using 12 workers ontwo CPU cores. Two sets of analysis were performed usinga sample acquired RF data of size 1152 (samples) x 104(elements). Computational time was calculated by increasingthe number of samples or increasing the interpolation gridsize and resolution. Benchmarking code can be made providedupon request.

D. Frame decimation

The accuracy and dynamic range of displacement trackinga 1.1 MHz ARF push is diminished. Therefore, to increasesensitivity and accuracy for real-time, intraprocedural targetingand monitoring, we can either increase transducer drivingpower or remove and decimate subsequent frames so thatinterframe displacement increases without changes to overallcumulative displacement. For every displacement map, RF ofdimension 104 elements x 1152 samples were stacked by thenumber of frames into a larger 2D array (104 x 80640 for a70 frames) and acquired at the highest frame-rate required toimage at a specific depth. To emphasize larger displacementestimates, we can down-sample a percentage of frames andfeed the resulting beamformed RF into the displacementestimation algorithm. In our optimization, we decimated bya factor of 0, 3, 5, 7, and 9 (i.e., removing zero, every 3rd,5th, 7th, or 9th frame). Therefore, for a 5 ms FUS pulse,the received 70 frames would be reduced to 23 frames afterdecimation by 3, allowing larger displacement measurementsbetween subsequent frames. Afterwards, SNR was calculatedon the displacement signals for every decimation factor.

E. Sample preparation

A polyacrylamide tissue-mimicking phantom with 4% agarpowder (Sigma-Aldrich, St. Louis, MO) as a scattering particle[47] was used to validate the displacement imaging. The phan-tom has an elastic modulus of 10 kPa, which was determinedusing the methodology described by Han et al. 2015 [48].

F. Human subject preparation

All human subjects (n=5) were recruited acquired in accor-dance with the 2013 Declaration of Helsinki. Ethical oversightand approval for this study was provided by the Institu-tional Review Board of Columbia University under protocolAAAR4475. Human subjects were seated comfortably withtheir forearm resting in a mechanical and movable cuff.The forearm was placed as to not introduce any physicalshifting during the experiment. Degassed ultrasound gel wasused to couple the transducer system bladder to the forearm.Compound B-mode imaging was used to initially place thetherapeutic transducer focus on the median nerve.

Heat stimulation pulses were delivered to the C6 dermatomeof the right arm (n = 1) using a custom-built thermofoil device.This initial feasibility study was set so that the FUS wasdelivered exactly when the heat signal was applied to the palm.FUS sonications (on target and off target) were randomizedwithin the 14 heat pulses delivered and the subject was askedto rate the intensity of the thermal stimulus using the Wong-Baker scale [49]. Though the scale ranges from 0 to 10, wekept thermal stimulation ratings in the range of 3 to 6.

G. Statistical analysis

All data was acquired in accordance with the ColumbiaUniversity Institutional Review Board regarding patient datacollection. Statistical testing was performed in Prism 8 (Graph-pad, San Diego, CA) using a non-parametric Mann-Whitneytest to test for differences between ratings given concurrentlywith FUS or sham sonication.

III. RESULTS

A. Signal-to-Noise Optimization

The imaging technique, using a 5 ms pulse sequence, inFig. 2 was optimized by calculating SNR during the FUSpulse in a gelatin tissue-mimicking phantom for decimateddisplacement estimation (0, 3, 5, 7, and 9 frames) at varyinglevels of focal pressures (Fig. 3). The displacement mapsshown are 0.16, 0.40, and 0.55 ms after triggering the FUSpulse. Displacements illustrate the ellipsoidal shape of theFUS beam but with increased lateral extent due to decimation.All calculations were performed using the ROI defined inFig. 3 (top), placed at the center of the transducer focus (30mm). Ten pressure levels, 5 realizations each, were used toacquire raw RF data using our technique. Post-hoc, frameswere decimated by a factor ranging from 0 to 9 frames andthe resulting displacement for each realization was estimatedusing normalized cross correlation. The resulting displacementmaps (Fig. 3 top) show larger interframe displacement valuesand illustrate the ellipsoidal focus better as more decimation





0.7 11.0Peak Rarefactional Pressure [MPa]

-20

-10

0

10

20

30

40

SN

R [

dB

]

2.0 3.4 4.5 5.7 6.9 7.9 8.9 10.00 1 2 3 4 5 6 7 8 9

Time [ms]

-2

-1

0

1

2

3

4

5

Dis

pla

cem

en

t [μ

m]

Decimation 0Decimation 3 Decimation 5 Decimation 7 Decimation 9

FUS

-5 0 5mm

20

22

24

26

28

30

32

34

mm

-5 0 5mm

-5 0 5mm

-5

-4

-3

-2

-1

0

1

2

3

4

5

Dis

pla

cem

en

t [μ

m]

Decimation 0 Decimation 5 Decimation 9

Fig. 3. Displacement tracking for decimation by 0, 3, 5, 7, and 9 frames in a tissue mimicking phantom. Top shows interframe displacement maps for 0,5, and 9 frame decimation. Bottom shows representative interframe displacement traces from the same RF data after decimation and corresponding SNRcalculations over pressures between 0.7 to 11 MPa peak negative.

is used. The associated interframe displacement traces areshown for a 5 ms FUS pulse in Fig. 3 bottom. Decimateddisplacements have fewer time-points within the FUS pulsebut have larger displacement values. Lastly, the SNR wascalculated within the pulse using the following equation [50],[51]:

SNR =µ

σ(1)

where µ is the mean displacement and σ is the varianceacross all realizations within the ROI. SNR was calculated fordisplacement values only within the FUS pulse. Results showthat, for a 5 ms pulse, decimation by 7 frames had the largestincrease in SNR by 15.09 ± 7.03 dB compared to no zerodecimation across all pressure levels. As such, displacementSNR increases with pressure, peaking at 5.7 MPa until thenoise from FUS harmonics severely impedes both B-modequality and the estimated displacement. Therefore, trade-offsbetween maximized SNR and number of frames within theFUS pulse indicate that optimal decimation rates should rangefrom 3 to 7 frames. Thus, for the following study, all RF datawas decimated by 5 frames before displacement estimation.

B. GPU Benchmarking

GPU parallelization of DAS beamforming using CUDAprogramming was compared to GPU parallelization using

MATLAB’s parallel computing toolbox, CPU parallelization(12 workers) over 2 cores, and CPU calculation on one core.Fig. 4 shows computation time over number of samples beam-formed. Averages over 3 realizations at increasing numberof RF samples show similar computation time for both GPUmethods until 105 samples. At > 105 samples, CUDA GPUperformance shows increasing computational speedup up to35 times the MATLAB GPU computational time. CUDA GPUmaintains consistent speed up over CPU and parallel CPU by300 and 60 times, respectively. Lastly, CUDA performs underreal-time criteria up to 105.6 samples.

The same computational time comparisons were performedfor different DAS interpolated grid size resolutions (1, 0.5,0.25, 0.125) for 70 frames of 1152 RF samples of base gridsize of 60 x 512 pixels. Fig. 5 shows CUDA DAS performswell compared to all other computational frameworks. TheMATLAB parallel toolbox and CUDA kernel performancewas similar until (240x2048 pixels) grid sizes. The CUDAkernel was not able to perform in real-time at grid sizes480x4096 pixels and above. Benchmarking results also showconsiderable speedup over CPU calculations.

C. Axial Displacement Focusing

The FUS beam was electronically focused in the axial direc-tion by calculating delays for each of the 4 annular ultrasonicelements so that the focal point is moved in the axial direction.





104 105 106

RF length (samples)

100-2

10-1

100

101

102

103

Co

mp

uta

tio

nal ti

me (

se

c)

CPU

Parallel CPU

MATLAB GPU

CUDA GPU

Fig. 4. Log plot of DAS computational time of CPU, parallel CPU, MATLABGPU parallel computing toolbox, and CUDA GPU over RF length.

60x512 120x1024 240x2048 480x4096Grid Size [px]

10-2

10-1

100

101

102

103

Co

mp

uta

tio

nal ti

me (

sec)

CPU

Parallel CPU

MATLAB Parallel GPU

CUDA GPU

Fig. 5. Computational time for CPU, parallel CPU, MATLAB parallel GPU,and CUDA GPU operations for increasing interpolated grid sizes for DASbeamforming. Computational time is expressed in log scale.

Variations in the beam-shape occurs due to the changes inthe f-number. Hydrophone measurements of the beam profilein free-field show axial focusing capabilities up to ±10 mmrelative to the geometric focal center. Axial focusing less than±5 mm leads to −3 dB drop off in pressure and greater than±5 mm shows greater drop off up to −6 dB. Displacementmaps in a homogeneous phantom were generated for eachfocal depth. Fig. 6 shows relevant focal displacements at -5,0, 5, and 10 mm focal depths relative to the geometric centerat the same time point. The figure illustrates maps showingmaximum interframe displacement, corresponding to the firstframes of FUS sonication (representative displacement trace isshown in Fig. 3). The ellipsoidal shape of the focus can bestbe visualized at ±5 mm. At larger axial focusing positions, thedisplacements succumb to unfilterable FUS interference noise

from overlapping FUS and imaging beams.

D. Human median nerve targeting

In Fig. 7 we implemented the pulse sequence to facilitatetargeting for human nerve neuromodulation in healthy subjects(n = 5). Nerve displacements up to 1 µm, overlapping withthe FUS focus, are shown in the second frame. Upwardsdisplacement/relaxation begins at 32 mm in frame 3 andpropagates outwards in frame 4. Interestingly, there are visibledifferences nerve displacement versus muscle (above 29 mm).This technique was used to first target the nerve, ensuringmaximum FUS delivery to the nerve. Varying levels of dis-placement for the five subjects were measured using a 5 msFUS pulse (5.6 MPa and 7.9 MPa peak rarefactional pressure).Mean displacement values at the center of the nerve (blackROI) from 7 FUS pulses are reported in Table I. Thoughthe focal pressure output from the transducer was the samefrom subject to subject, the amount of displacement variedfrom 10 to 30 microns in peak cumulative displacement.After targeting validation, displacement images were used tomonitor neuromodulation.

TABLE ISUMMARY OF MEASURED DISPLACEMENTS ACROSS SUBJECTS.

Subject Pressure[MPa]

Interframedisplacement[µm]

Cumulativedisplacement[µm]

1 5.6 2.1±0.3 18.3±2.42 5.6 5.5±0.3 50.1±3.43 7.9 4.2±0.3 31.3±2.34 7.9 5.2±0.5 42.3±5.55 7.9 5.1±1.6 40.7±10.7

A neuromodulation experiment was conducted to validatethe technique in a sensory neuromodulation experiment. Fif-teen 2 second thermal pulses were delivered to the C6 der-matome of the human palm at a random interval between 3and 4 minutes and a subject (n = 1) was asked to rate theintensity of the pulse based on the Wong-Baker scale [49].FUS (7.9 MPa rarefactional pressure) and sham (no FUS andoff target FUS) stimulations were randomized among the 15heat pulses delivered in a single trial. Displacement imagingallowed monitoring of all FUS pulses and measurement ofdisplacement at the nerve. Fig. 8 shows one frame of thesupplementary video 1, indicating FUS engagement of themedian nerve. To investigate acute effects of FUS on sensoryperception, a pulse duration of 5 ms was chosen to ensureFUS was applied during sensory stimulus conduction; basedon the average conduction velocity in a healthy human subject[52] and the approximate distance of the FUS focus to thethermal stimulus (approximately 5 - 6 cm). Both interframeand cumulative displacement were estimated during the FUSpulse transmission at the peak of temperature delivery. Forthis particular subject, the peak interframe and cumulativedisplacements were estimated at the center of the nerve (blackROI) during neuromodulation was 5.1 ± 0.7 and 40.7 ± 7.4microns, respectively. The maximum interframe displacementwas achieved at the beginning of the pulse and the peakcumulative displacement was acquired at the end of the FUS





-5mm0 5

0 mm

-5 0mm

5

5 mm

-5 0mm

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

Dis

pla

cem

en

t [μ

m]

5

10 mm25

30

35

40

mm

-5 mm

-5mm0 5

Fig. 6. Displacement maps produced using a gelatin tissue-mimicking phantom at -5, 0, 5, and 10 mm focal depths. Frames shown are at the peak interframedisplacement for a 5 ms FUS push.

-5 0 5mm

20

25

30

35

mm

-5 0 5mm

-5 0 5mm

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

Dis

pla

cem

en

t [μ

m]

-5 0 5mm

t = 0.00 ms t = 0.71 ms t = 1.72 ms t = 3.57 ms

- - Median Nerve - ROI

Fig. 7. Representative tracked displacements before, during, and after FUS push in the human subject forearms. The Median nerve is outlined and centeredat 30 mm. The median nerve was outlined based on B-mode images where the center of the nerve was identified and an area of 19.6 mm2 was selectedencompassing a radius of 2.5 mm that corresponds to the cross-sectional diameter of the median nerve.

pulse. Sources of error in the displacement curve may be dueto slight movement in the subjects’ arm during the procedure.Preliminary data indicates that FUS may change subjectivethermal perception by modulating the median nerve duringheat stimulation. Fig. 9 shows ratings from heat stimuli withFUS vs sham. A 0.9643 pain rating unit decrease, without sig-nificance, was found in FUS sonications vs. sham stimulations(p = 0.0547; two-tailed unpaired Mann-Whitney test) wherelower ratings were indicated for lower needle-like thermalpain.

For further validation of our technique to neuromodulation,we note differences between FUS delivery when FUS isapplied directly to the nerve vs off-target. Fig. 10 illustrates arepresentative displacement image when the focus was off themedian nerve (white outline). FUS displacement still appearsat the focus of the transducer, however the nerve was posi-tioned 5 mm away. As a result, the displacement at the centerof the nerve (black ROI) had a peak cumulative displacementof 3.2±1.7 microns when the focus was off-target. Moreover,

the peak interframe displacement was 1.7±0.3 microns at theend of the FUS pulse rather than the beginning as in Fig. 8 witha 5 µm difference in peaks. Finally, the thermal stimulationswhich were off-target had an average rating of 1 higher thanwhen FUS was acting on the nerve directly (p = 0.4524; two-tailed unpaired Mann-Whitney test).

IV. DISCUSSIONUsing a confocally aligned imaging and FUS transducer,

we can perform real-time displacement tracking, mediatedby CUDA-accelerated DAS, and axial focal steering using asingle ultrasound DAQ system without interleaving. Half ofthe channels can be used to drive the FUS-guided system withsimultaneous custom waveforms programmed for imaging onhalf the channels and FUS on the other. A major advantageof using this technique, is the ability to perform RF stackingand displacement tracking in real-time.

Using the Cramer-Rao lower bound [26] for a SNR of 30dB, a correlation coefficient of 0.98, and a window length





- - Median Nerve

-5 0 5

mm

20

22

24

26

28

30

32

34

mm

-10

-5

0

5

10

Dis

pla

cem

en

t [μ

m]

- ROI

t = 0.86 ms

0 1 2 3 4 5 6 7 8 9

-5

0

5

Inte

rfra

me D

isp

lacem

en

t [μ

m]

Time [ms]

0

20

40

60

Cu

mu

lati

ve D

isp

lacem

en

t [μ

m]

FUS

Fig. 8. Illustration of targeting and monitoring FUS neuromodulation using simultaneous displacement imaging. Left shows a displacement map at 0.86 msafter triggering the FUS pulse. The median nerve is located at a depth of approximately 31 mm and was outlined based on B-mode images using an areaof 19.6 mm2. Right shows average and standard deviation interframe and cumulative displacement traces over 7 FUS on-target stimulations from the ROIshown in black.

FUS Sham

0

2

4

6

8

Th

erm

al

Ra

tin

g

p = 0.0547

Fig. 9. Thermal ratings for heat pulses with FUS and sham treatment to themedian nerve (p = 0.0547, two-tailed unpaired Mann-Whitney test)

of 11.25 µs our technique can ideally achieve displacementsabove 0.6711 µm where they succumb to jitter. Though,our transducer can achieve ±10 mm axial steering of theFUS focus, there is a drop off in tracked displacement. Thisphenomenon is especially apparent at lower positions, wherethese low pressure pushes are below the noise floor. However,the ±5 mm range is more than adequate for targeting themedian nerve in the human arm and accounting for small

movements, limiting off-target effects. In the distal half ofthe forearm, the variation in depth of the median nerve iswell within the range of 5 mm. Furthermore, since activecompounded B-mode imaging is continuously used betweendisplacement mapping sonications, we can actively account forany movement during the entire procedure.

Benchmarking clearly shows the advantages of parallelGPU computing for conventional DAS compared to CPUcalculations. However, the advantage of CUDA beamformingover using GPU matrices in the MATLAB toolbox are notas easily distinguishable. Computational speedup only occursat high data volumes and for displacement mapping, higherresolution interpolation grids increase SNR of displacementsand may benefit from CUDA operation. This deviation maybe due to how CUDA-written programs efficiently use sharedmemory whereas the MATLAB parallel computing toolboxdoes not. Other techniques, such as Doppler functional Ultra-Sound (fUS) would also benefit from utilizing CUDA kernelsas compounding more than 200 frames can become compu-tationally intensive. Not included in the above benchmarkinganalysis is the acceleration of the initialization of the imagingsequence. Because the MATLAB toolbox technique involvespre-indexing and delay calculation, initialization of scriptsusing large data volumes may take minutes to initialize.However, CUDA calculation is performed in real-time, lead-ing to less wait time and faster in-procedure operation. Onthe other hand, as of now, the proposed technique achieves





-5 0 5

mm

20

22

24

26

28

30

32

34

mm

Dis

pla

cem

en

t [μ

m]

t = 0.86 ms

-10

-5

0

5

10

- - Median Nerve - ROI

0 1 2 3 4 5 6 7 8 9

Time [ms]

FUS

On Target Off Target

0

2

4

6

8

Pa

inR

ati

ng

- On Target- Off Target

-5

0

5

Inte

rfra

me D

isp

lacem

en

t [u

m]

Fig. 10. Representative off-target stimulation. Left shows the displacement map for a FUS pulse where the nerve was located 5 mm laterally from the FUSfocus (off target). Right shows the interframe displacement at the ROI on the nerve for on- vs off-target stimulations and corresponding thermal ratings foreach trace.

real-time specifications and CUDA beamforming may benefitfrom calculating delays before initialization instead of duringacquisition so that during imaging operation, other computa-tionally intensive operations may take priority, speeding upthe technique. This technique is feasible to implement inother more conventional ultrasound systems if an external FUStransducer is used or a custom device to allocate channels toa FUS transducer and imaging transducer, given the sufficientframe rates (above 10 kHz) and output power (0.5 to 1MPa) can be achieved. Lower number of imaging channelswill not impede displacement mapping, as probes such asthe P4-2 (64 channels) can be used to track tissue motion[48]. Lastly, the imaging frequency must be selected so asto minimize the overlap of the FUS transducer center andharmonic frequencies.

We demonstrate our technique’s capability to facilitate tar-geting (n = 5) of the median nerve in and neuromodulation (n= 1) in healthy human subjects. Displacement maps show thatthe intensity of the displacement increases as the pressure isincreased. Due to the heterogeneity of the forearm and bound-ary effects between muscle, connective tissue, and nerve fibers,the displacement map does not necessarily follow the expectedellipsoidal focus geometry. Additionally, the variance withina single subject was consistent between FUS pulses (maxstandard deviation of 10.7 microns). However, our techniquereveals that same output pressures do not necessarily translateto the same FUS modulation efficiency (nerve displacement)

between subjects. We measured 50 µm nerve displacementfrom a 5.6 MPa, 5 ms pulse in 1 subject but 18 µm inanother from the same pulse parameters. This can be dueto a number of factors: location of the transducer on theforearm, the incident angle to the skin, tissue properties, andhow well coupled the transducer system is to the skin. Finally,we present a preliminary findings showing FUS effects onthermal pain perception. Results show that thermal pulseswith coincident FUS sonication had lower thermal ratingsthan sham pulses. Moreover, pulses that were off-target to thenerve had a higher rating than when the focus was positionedat the center of the nerve. Displacement maps show thatdisplacement characteristics are vastly different between on-and off-target pulses. The peak of interframe displacementwas near the beginning of the pulse for on-target versus theend of the pulse for off-target (Fig. 10). FUS was capableof suppressing pain signals only when directly targeted tothe nerve trunk. This effect could be explained by either thegeneration of afferent signals at the nerve trunk, generatinga masking effect of pain, changing how the subject perceivespain or, most likely, by the direct suppression or interruptionof signals by FUS at a more proximal portion of the nervetrunk. The masking effect is unlikely to occur because thesonication using the parameters used in this study withoutany other stimuli did not generate any sensory responseitself. Future studies will explore this hypothesis, for example,targeting receptive fields at the skin in order to generate tactile





sensations during a concurrent pain task.Limitations of operating at 1.1 MHz as opposed to higher

frequencies, includes lower tissue displacements as the tis-sue absorption and acoustic radiation force decrease withfrequency. Interleaving transmits would lower the amplitudeof displacements further, thus simultaneous imaging withinthe pulse is essential. The trade-off of this technique is theinterference from the FUS transducer. Ideally, the transducerbandwidths should be separated as far as possible to reducethe interference, but the depth penetration limits the rangeof frequencies. Additional filtering can minimize interference,but an inherent smoothing of the RF data is inevitable.Nonetheless, we can adapt this single DAQ method to otherelastography techniques such as harmonic motion imaging(HMI) by amplitude modulating the HIFU [42]. As notedbefore, hydrophone measurements show that we emitted apseudo-CW push due to electronic actuation of elements. Theeffect of this timing on displacement has yet to be determined.However, more complex coded excitations can be developedto achieve true CW such as the ones presented by Tiran et al.2015 [53].

In ultrasound neuromodulation literature thus far, propertargeting is an often overlooked aspect of experimental design,but having real-time feedback and confirmation of targetingcan lead to more conclusive results. Crucially, we foundthat transducer driving pressure does not necessarily lead toreproducible therapeutic levels subject to subject, providing anexplanation for the wide variety of US pressures reported inneuromodulation, i.e., 1.8 MPa, 3.2 MPa, or even 50 MPato achieve an action potential [54]. Displacement imagingreveals real-time feedback on how much FUS is engagingthe nerve, which may vary subject to subject or in the samesubject depending on the coupling condition (i.e. coupling gel,incidence angle). We show that transducer focal pressure maynot be a valid indication of FUS nerve modulation efficacy;using displacement imaging may facilitate comparison to otherreports of FUS neuromodulation. Furthermore, the ability tofocus-shift adds to the ability to modify targeting during theprocedure in remedy of poor coupling, movement artifacts, andoff-target effects. Lastly, we have shown that variation can bedetected and imaged with this technique which may play animportant role in further characterizing the mechanism of FUSneuromodulation.

V. CONCLUSION

By deriving a new pulse sequence and hardware combina-tion to simultaneously drive a FUS and imaging transducerwith a single ultrasound DAQ system, we have shown inthis article that tracking displacements within the FUS pulsefor targeting and monitoring can be accomplished. The pulsesequence and frame-rate were experimentally optimized usinga homogeneous tissue-mimicking phantom and applied toin vivo human median nerve stimulation. Furthermore, thetechnique is able to operate in real-time and account for shiftsin targeting using axial beam steering. We found that themicron displacements in the nerve were able to confirm FUSdelivery. Lastly, the imaging technique presented herein was

successfully validated in an experiment demonstrating FUSeffects on thermal perception where the subject experienceda 0.9643 pain rating unit decrease in the needle-like painsensation. Moreover, our technique can validate on- and off-targeting of the nerve indicating 5-micron differences at thetransducer driving pressures explored in this study. The imag-ing technique developed here is not restricted to the tissuesvalidated in this article, but can be applied to any ultrasound-accessible soft tissue found elsewhere in the body, such asthe brain. Furthermore, current studies would benefit from amethod that allows for real-time targeting, allowing for greaterconfidence in the results of FUS mechanistic studies.

ACKNOWLEDGMENTSThis work was supported by the Defense Advanced Re-

search Projects Agency (DARPA) Biological Technologies Of-fice (BTO) Electrical Prescriptions (ElectRx) (Grant/ContractNo. DARPA HR0011-15-2-0054), the National Institute ofBiomedical Imaging and Bioengineering of the National Insti-tutes of Health under Award Number R01EB027576, Sound-Stim Inc, and Google X.

REFERENCES[1] O. Naor, S. Krupa, and S. Shoham, “Ultrasonic neuromodulation,”

Journal of Neural Engineering, vol. 13, no. 3, p. 031003, may 2016.[2] A. Fomenko, C. Neudorfer, R. F. Dallapiazza, S. K. Kalia, and A. M.

Lozano, “Low-intensity ultrasound neuromodulation: An overview ofmechanisms and emerging human applications,” Brain Stimulation,vol. 11, no. 6, pp. 1209 – 1217, 2018.

[3] C. Pasquinelli, L. G. Hanson, H. R. Siebner, H. J. Lee, and A. Thielscher,“Safety of transcranial focused ultrasound stimulation: A systematicreview of the state of knowledge from both human and animal studies,”Brain Stimulation, 2019.

[4] R. L. King, J. R. Brown, W. T. Newsome, and K. B. Pauly, “Effectiveparameters for ultrasound-induced in vivo neurostimulation,” Ultrasoundin Medicine and Biology, vol. 39, no. 2, pp. 312–331, 2013.

[5] H. Baek, K. J. Pahk, and H. Kim, “A review of low-intensity focused ul-trasound for neuromodulation,” Biomedical Engineering Letters, vol. 7,no. 2, pp. 135–142, 2017.

[6] J. Kubanek, J. Shi, J. Marsh, D. Chen, C. Deng, and J. Cui, “Ultrasoundmodulates ion channel currents,” Scientific Reports, vol. 6, p. 24170,2016.

[7] H. A. S. Kamimura, S. Wang, H. Chen, Q. Wang, C. Aurup, C. Acosta,A. A. O. Carneiro, and E. E. Konofagou, “Focused ultrasound neuro-modulation of cortical and subcortical brain structures using 1.9 MHz,”Medical Physics, vol. 43, no. 10, pp. 5730–5735, sep 2016.

[8] Y. Tufail, A. Matyushov, N. Baldwin, M. L. Tauchmann, J. Georges,A. Yoshihiro, S. I. Tillery, and W. J. Tyler, “Transcranial PulsedUltrasound Stimulates Intact Brain Circuits,” Neuron, vol. 66, no. 5,pp. 681–694, 2010.

[9] W. Lee, H.-C. Kim, Y. Jung, Y. A. Chung, I.-U. Song, J.-H. Lee, and S.-S. Yoo, “Transcranial focused ultrasound stimulation of human primaryvisual cortex,” Scientific Reports, vol. 6, no. 1, p. 34026, 2016.

[10] P. P. Lele, “Effects of focused ultrasonic radiation on peripheral nerve,with observations on local heating,” Experimental Neurology, vol. 8,no. 1, pp. 47–83, 1963.

[11] W. J. Tyler, Y. Tufail, M. Finsterwald, M. L. Tauchmann, E. J. Olson, andC. Majestic, “Remote excitation of neuronal circuits using low-intensity,low-frequency ultrasound,” PLoS ONE, vol. 3, no. 10, 2008.

[12] H. A. S. Kamimura, A. Conti, N. Toschi, and E. E. Konofagou, “Ultra-sound neuromodulation: Mechanisms and the potential of multimodalstimulation for neuronal function assessment,” Frontiers in Physics,vol. 8, p. 150, 2020.

[13] C. J. Wright, J. Rothwell, and N. Saffari, “Ultrasonic stimulation ofperipheral nervous tissue: An investigation into mechanisms,” Journalof Physics: Conference Series, vol. 581, no. 1, p. 012003, 2015.

[14] H. Kim, S. Taghados, and K. Fischer, “Noninvasive transcranial stim-ulation of rat abducens nerve by focused ultrasound,” Ultrasound inmedicine & biology, vol. 38, no. 9, pp. 1568–1575, 2012.





[15] M. E. Downs, S. A. Lee, G. Yang, S. Kim, Q. Wang, and E. E.Konofagou, “Non-invasive peripheral nerve stimulation via focusedultrasound in vivo,” Physics in Medicine and Biology, vol. 63, no. 3,p. 035011, jan 2018.

[16] W. J. Elias, D. Huss, T. Voss, J. Loomba, M. Khaled, E. Zadicario, R. C.Frysinger, S. A. Sperling, S. Wylie, S. J. Monteith et al., “A pilot studyof focused ultrasound thalamotomy for essential tremor,” New EnglandJournal of Medicine, vol. 369, no. 7, pp. 640–648, 2013.

[17] W. Legon, T. F. Sato, A. Opitz, J. Mueller, A. Barbour, A. Williams, andW. J. Tyler, “Transcranial focused ultrasound modulates the activity ofprimary somatosensory cortex in humans,” Nature neuroscience, vol. 17,no. 2, p. 322, 2014.

[18] W. Lee, H. Kim, Y. Jung, I.-U. Song, Y. A. Chung, and S.-S. Yoo,“Image-guided transcranial focused ultrasound stimulates human pri-mary somatosensory cortex,” Scientific reports, vol. 5, p. 8743, 2015.

[19] T. C. Dickey, R. Tych, M. Kliot, J. D. Loeser, K. Pederson, and P. D.Mourad, “Intense focused ultrasound can reliably induce sensations inhuman test subjects in a manner correlated with the density of theirmechanoreceptors,” Ultrasound in medicine & biology, vol. 38, no. 1,pp. 85–90, 2012.

[20] L. Gavrilov, “Use of focused ultrasound for stimulation of nerve struc-tures,” Ultrasonics, vol. 22, no. 3, pp. 132–138, 1984.

[21] B. Yoshi and B. H. S. L. E. K. E. Lumpkin, “Focused ultrasound excitesaction potentials in mammalian peripheral neurons in intact tissue,”PNAS (in review), 2019.

[22] F. Wu, Z.-B. Wang, H. Zhu, W.-Z. Chen, J.-Z. Zou, J. Bai, K.-Q. Li, C.-B. Jin, F.-L. Xie, and H.-B. Su, “Feasibility of us-guidedhigh-intensity focused ultrasound treatment in patients with advancedpancreatic cancer: initial experience,” Radiology, vol. 236, no. 3, pp.1034–1040, 2005.

[23] L. H. Treat, N. McDannold, N. Vykhodtseva, Y. Zhang, K. Tam, andK. Hynynen, “Targeted delivery of doxorubicin to the rat brain attherapeutic levels using mri-guided focused ultrasound,” Internationaljournal of cancer, vol. 121, no. 4, pp. 901–907, 2007.

[24] K. Hynynen, “Mri-guided focused ultrasound treatments,” Ultrasonics,vol. 50, no. 2, pp. 221–229, 2010.

[25] J. T. de Bever, H. Odéen, L. W. Hofstetter, and D. L. Parker, “Simul-taneous MR thermometry and acoustic radiation force imaging usinginterleaved acquisition,” Magn. Reson. Med., vol. 79, no. 3, pp. 1515–1524, mar 2018.

[26] W. F. Walker and G. E. Trahey, “A Fundamental Limit on DelayEstimation Using Partially Correlated Speckle Signals,” IEEE Trans.Ultrason. Ferroelectr. Freq. Control, vol. 42, no. 2, pp. 301–308, 1995.

[27] J. Strasser, M. T. Haindl, R. Stollberger, F. Fazekas, and S. Ropele,“Magnetic resonance elastography of the human brain using a multi-phase dense acquisition,” Magnetic resonance in medicine, 2019.

[28] J. Vappou, P. Bour, F. Marquet, V. Ozenne, and B. Quesson, “MR-ARFI-based method for the quantitative measurement of tissue elasticity:Application for monitoring HIFU therapy,” Phys. Med. Biol., vol. 63,no. 9, p. 095018, may 2018.

[29] E. A. Kaye, J. Chen, and K. B. Pauly, “Rapid MR-ARFI method forfocal spot localization during focused ultrasound therapy,” Magn. Reson.Med., vol. 65, no. 3, pp. 738–743, mar 2011.

[30] N. McDannold and S. E. Maier, “Magnetic resonance acoustic radiationforce imaging,” Med. Phys., vol. 35, no. 8, pp. 3748–3758, 2008.

[31] J.-L. Gennisson, T. Deffieux, M. Fink, and M. Tanter, “Ultrasoundelastography: principles and techniques,” Diagnostic and interventionalimaging, vol. 94, no. 5, pp. 487–495, 2013.

[32] C. Maleke and E. E. Konofagou, “In vivo feasibility of real-timemonitoring of focused ultrasound surgery (fus) using harmonic motionimaging (hmi),” IEEE Transactions on biomedical Engineering, vol. 57,no. 1, pp. 7–11, 2009.

[33] S. Y. Lee, A. R. Cardones, J. Doherty, K. Nightingale, and M. Palmeri,“Preliminary results on the feasibility of using arfi/swei to assesscutaneous sclerotic diseases,” Ultrasound in medicine & biology, vol. 41,no. 11, pp. 2806–2819, 2015.

[34] J. D’hooge, A. Heimdal, F. Jamal, T. Kukulski, B. Bijnens, F. Rade-makers, L. Hatle, P. Suetens, and G. R. Sutherland, “Regional strain andstrain rate measurements by cardiac ultrasound: principles, implemen-tation and limitations,” European Journal of Echocardiography, vol. 1,no. 3, pp. 154–170, 2000.

[35] P. Bihari, A. Shelke, T. Nwe, M. Mularczyk, K. Nelson, T. Schmandra,P. Knez, and T. Schmitz-Rixen, “Strain measurement of abdominal aorticaneurysm with real-time 3d ultrasound speckle tracking,” EuropeanJournal of Vascular and Endovascular Surgery, vol. 45, no. 4, pp. 315–323, 2013.

[36] K. Nightingale, D. Stutz, R. Bentley, and G. Trahey, “Acoustic radiationforce impulse imaging: EX vivo and in vivo demonstration of transientshear wave propagation,” in Proceedings - International Symposium onBiomedical Imaging, vol. 2002-Janua, no. 2. Elsevier, feb 2002, pp.525–528.

[37] K. Nightingale, “Acoustic Radiation Force Impulse (ARFI) Imaging: AReview,” Current Medical Imaging Reviews, vol. 7, no. 4, pp. 328–339,nov 2011.

[38] M. Palmeri, M. Wang, J. Dahl, K. Frinkley, and K. Nightingale,“Quantifying Hepatic Shear Modulus In Vivo Using Acoustic RadiationForce,” Ultrasound in Medicine & Biology, vol. 34, no. 4, pp. 546–558,apr 2008.

[39] G. F. Pinton, J. J. Dahl, and G. E. Trahey, “Rapid tracking of smalldisplacements using ultrasound,” Proceedings - IEEE Ultrasonics Sym-posium, vol. 4, no. 6, pp. 2062–2065, 2005.

[40] Y. Han, S. Wang, H. Hibshoosh, B. Taback, and E. Konofagou, “Tumorcharacterization and treatment monitoring of postsurgical human breastspecimens using harmonic motion imaging (HMI),” Breast CancerResearch, vol. 18, no. 1, p. 46, 2016.

[41] R. Xia and A. K. Thittai, “Real-Time Monitoring of High-IntensityFocused Ultrasound Treatment Using Axial Strain and Axial-ShearStrain Elastograms,” Ultrasound in Medicine and Biology, vol. 40, no. 3,pp. 485–495, mar 2014.

[42] C. Maleke and E. E. Konofagou, “In vivo feasibility of real-timemonitoring of focused ultrasound surgery (FUS) using harmonic motionimaging (HMI),” IEEE Transactions on Biomedical Engineering, vol. 57,no. 1, pp. 7–11, jan 2010.

[43] S. A. Lee, H. A. S. Kamimura, M. T. Burgess, and E. E. Konofagou,“Displacement imaging for focused ultrasound peripheral nerve neu-romodulation,” IEEE Transactions on Medical Imaging, pp. 1–1, may2020.

[44] J. S. Jeong, J. H. Chang, and K. K. Shung, “Ultrasound transducerand system for real-time simultaneous therapy and diagnosis for non-invasive surgery of prostate tissue.” IEEE transactions on ultrasonics,ferroelectrics, and frequency control, vol. 56, no. 9, pp. 1913–22, sep2009.

[45] M. Fink, Bercoff, and Tanter, “Supersonic Shear Imaging : A NewTechnique for Soft Tissue Elasticity Mapping,” IEEE transactions onultrasonics, ferroelectrics, and frequency control, vol. 51, no. 4, pp.396–409, 2004.

[46] J. Luo and E. Konofagou, “A fast normalized cross-correlation calcula-tion method for motion estimation,” IEEE transactions on ultrasonics,ferroelectrics, and frequency control, vol. 57, no. 6, pp. 1347–57, 2010.

[47] J. Vappou, G. Y. Hou, F. Marquet, D. Shahmirzadi, J. Grondin, andE. E. Konofagou, “Non-contact, ultrasound-based indentation methodfor measuring elastic properties of biological tissues using harmonicmotion imaging (hmi),” Physics in Medicine & Biology, vol. 60, no. 7,p. 2853, 2015.

[48] Y. Han, G. Y. Hou, S. Wang, and E. Konofagou, “High intensity fo-cused ultrasound (HIFU) focal spot localization using harmonic motionimaging (HMI),” Phys. Med. Biol., vol. 60, no. 15, pp. 5911–5924, aug2015.

[49] D. L. Wong and C. M. Baker, “Pain in children: comparison ofassessment scales,” Pediatr Nurs, vol. 14, no. 1, pp. 9–17, 1988.

[50] A. H. Dhanaliwala, J. A. Hossack, and F. W. Mauldin, “Assessingand improving acoustic radiation force image quality using a 1.5-Dtransducer design,” IEEE Trans. Ultrason. Ferroelectr. Freq. Control,vol. 59, no. 7, pp. 1602–1608, 2012.

[51] V. Shamdasani and Y. Kim, “Two-dimensional autocorrelation methodfor ultrasound-based strain estimation,” in Annu. Int. Conf. IEEE Eng.Med. Biol. - Proc., vol. 26 II, 2004, pp. 1380–1383.

[52] V. K. Nielsen, “Sensory and motor nerve conduction in the median nervein normal subjects,” Acta Med. Scand., vol. 194, no. 1-6, pp. 435–443,apr 2009.

[53] E. Tiran, T. Deffieux, M. Correia, D. Maresca, B.-F. Osmanski, L.-A.Sieu, A. Bergel, I. Cohen, M. Pernot, and M. Tanter, “Multiplane waveimaging increases signal-to-noise ratio in ultrafast ultrasound imaging,”Physics in Medicine & Biology, vol. 60, no. 21, p. 8549, 2015.

[54] J. Blackmore, S. Shrivastava, J. Sallet, C. R. Butler, and R. O. Cleveland,“Ultrasound Neuromodulation: A Review of Results, Mechanisms andSafety,” pp. 1509–1536, jul 2019.





Stephen A. Lee received the bachelor’s degree fromthe University of North Carolina at Chapel Hill,Chapel Hill, NC, USA in 2015, and the Master’sof Science degree from Columbia University, NewYork, NY, USA in 2018. In 2016, he enrolledin the M.S./Ph.D. Program with the Departmentof Biomedical Engineering, Columbia University,New York, NY, USA where he is currently work-ing towards his Ph.D. His current research inter-ests include studying the mechanism of ultrasoundneuromodulation and how ultrasound imaging and

simulation techniques can be used understand this phenomenon. In 2019, hewas the recipient of the IEEE IUS student paper competition award.

Hermes A.S. Kamimura (Member, IEEE) receivedthe B.S. degree in medical physics and the M.S. andPh.D. degrees in physics applied to medicine andbiology from the University of São Paulo, RibeirãoPreto, Brazil, in 2008, 2011, and 2016, respectively.He conducted research projects in therapeutic andultrasound imaging at Mayo Clinic, Rochester, MN,USA, in 2010, and also at Columbia University,New York, NY, USA, in 2014, in student exchangeprograms. He was a Postdoctoral Research Scientistwith French Alternative Energies and Atomic En-

ergy Commission (CEA), Gif-sur-Yvette, France, and also with ColumbiaUniversity, where he is currently an Associate Research Scientist. His researchinterests include harmonic motion imaging and therapeutic ultrasound span-ning both ultrasound neuromodulation and ultrasound mediated bloodbrainbarrier disruption for targeted brain drug delivery. Dr. Kamimura is a mem-ber of the Brazilian Physical Society and Brazilian Society of BiomedicalEngineering. He was a recipient of the Outstanding Reviewer Award forPhysics in Medicine and Biology, IOP Publishing in 2018, and the Best Ph.D.Dissertation Award in the field of medical physics in 2016 by the BrazilianPhysical Society. He serves as a Topic Editor for Frontiers in Physics andFrontiers in Physiology.

Elisa E. Konofagou (S98A99M03) is currently theRobert and Margaret Hariri Professor of biomedicalengineering and a Professor of radiology as well asthe Director of the Ultrasound and Elasticity Imag-ing Laboratory with the Biomedical EngineeringDepartment, Columbia University, New York, NY,USA. She has co-authored over 170 peer-reviewedjournal articles. Her current research interests in-clude the development of novel elasticity imag-ing techniques and therapeutic ultrasound methodsand more notably, myocardial elastography, elec-

tromechanical and pulse wave imaging, harmonic motion imaging, focusedultrasound therapy, and drug delivery in the brain, with several clinicalcollaborations in the Columbia Presbyterian Medical Center, New York, NY,USA, and elsewhere. Prof. Konofagou was a recipient of awards, such as theCAREER Award from the National Science Foundation and the Nagy Awardfrom the National Institutes of Health as well as others from the AmericanHeart Association, the Acoustical Society of America, the American Instituteof Ultrasound in Medicine, the Wallace H. Coulter Foundation, the BodossakiFoundation, the Society of Photo-Optical Instrumentation Engineers, andthe Radiological Society of North America. She is a Technical CommitteeMember of the Acoustical Society of America, the International Societyof Therapeutic Ultrasound, the IEEE Engineering in Medicine and BiologyConference, the IEEE International Ultrasonics Symposium, and the AmericanAssociation of Physicists in Medicine. She serves as an Associate Editor forthe IEEE Transactions on ultrasonics, ferroelectrics, and frequency control,Ultrasonic Imaging, and Medical Physics.


Documents

Displacement imaging during focused ultrasound median ... · [3]. Ultrasound has been proven capable of modulating ion channel, peripheral nerve, and brain activity in a multitude